[hpeterson]

On version 2.1.1 I have a HUGE image file we are trying to process on one of our servers. The Hierachical message processor fails, runs out of memory. I stop the pid for the processing engine attached to it and restart it and the HME. The HME fails again. I won't let go and fail even if I run core_servers stop and start. I don't want to run the XIB clean program. What do I need to do to get clear the message?

[Rhenus ICT Servi...]

Aha....this is one of those nasty things that is rather cumbersome in XIB; when we face problems like this we stop the processes from the XIB Task Monitor (not using the "core_servers stop" command) and then we DISABLE the INTEGRATION or ACTIVITY that causes the problem. When you start all processes in the Task Monitor again the message that is causing the problem should go into ERROR in the messagelog and the processing is stopped.

Wouldn't it be nice if you could just select the troublesome message in the messagelog with the right-mouse-button and select an option like 'Stop processing message'...but that's just wishfull thinking.

Good Luck!

[Sairam Aripirala]

Hi, You can actually do one thing, if you are sure about the HME being used for this SQA open Queuemonitor from sysman, check the message in particular HME Queue. On right hand side it shows up the message details. You can right click on the entry and see the activity or integration. Select it and say delete, you will have the option. Try to stop and start corresponding PE, it goes into error state and other massages would get processed.

[Kavindra Kulathi...]

sometimes it tends to halt when you ahve tounsadof messages in queues, i do it like this in production environemnt

1. Stop all tasks in task monitor
2. start only the queue process
3. Open to queue manger and find the troubling entry/entired and save the logger ID/ID's of that queue entries and remove it from queue
4. user logger_utility to inactivate the logger respective entry/entries
5. restart all processes.
6. Analyze the log entry for failure and resend it back once the problem is identified

i have been using this in production env for some time and was able to recover system in most critical situations without loosing any messages in queue, this applyes to XIB 2.1.1 with SP3